Sovereign AI for Capital Markets
Mickai is a sovereign intelligence operating system that regulated businesses own and run inside their own walls. For capital markets, that means artificial intelligence that never leaves the perimeter, with proof that survives audit, discovery, and time.
The market the public cloud cannot lawfully reach
Capital markets run on data that cannot leave the building. Order flow, client positions, model risk documentation, insider lists, and surveillance records are among the most tightly governed information a regulated firm holds. The prevailing generation of enterprise artificial intelligence assumes the opposite: that data can be shipped to a hyperscale region, processed on shared infrastructure, and returned. For a bank, a broker dealer, an asset manager, or an insurer operating under formal supervision, that assumption is often not a preference to be weighed. It is a line that cannot be crossed.
We built for the firms on the wrong side of that line. Our position is simple. We provide intelligence that never leaves the perimeter, and we prove it in a way that survives audit, discovery, and time.
What we run and where
We run on the customer's own hardware, on premises and air gapped, with zero data egress and no public cloud round trip. Nothing is sent out to be scored, and nothing depends on a remote endpoint staying available. The intelligence is resident, and it stays resident.
Inside that perimeter we run about fifty specialist models, twenty five domain and twenty five operational, with cross model routing under a deterministic arbiter. The arbiter matters more than any single model. It means outputs are reproducible: the same inputs, under the same policy, produce the same result, which is the baseline a supervised firm needs before it can put a model anywhere near a regulated decision. We refer to these as our sovereign models throughout. What they are trained to do is what the buyer cares about, not their provenance.
The capability arrives as studios, each a focused surface for a specific job. Nemesis handles fraud and anti money laundering. Plutus covers finance and FP&A. Tyche runs underwriting. Prometheus does forecasting. Iris handles customer service. Nomos covers compliance, Astraea covers legal, Panacea covers clinical work, Pythia delivers business intelligence, and Aletheia performs audit. Vinis provides voice, the Agentic Marketing Team drives outbound work, and Trust Agent holds the perimeter. For a trading business, that maps directly onto the desks that already exist: surveillance and AML through Nemesis, treasury and planning through Plutus, book level forecasting through Prometheus, and continuous trade surveillance evidence through Aletheia, all inside one governed system rather than a patchwork of external services.
Proof that does not depend on trusting us
A supervised firm cannot run on assurances. It needs evidence. The Open Audit Record is how we produce it. Every consequential action is signed under post-quantum cryptography (FIPS 204 ML-DSA-65, with ML-KEM-768) and hash chained into a tamper evident, append only ledger. Anyone can verify that ledger offline, for decades, without trusting the vendor and without a network connection.
For capital markets that is the difference between a model you have to defend and a model that defends itself. When a supervisor, an internal auditor, or opposing counsel asks what the system did on a given day, the record answers, and the answer cannot have been quietly rewritten after the fact. We also offer this as OAR-as-a-Service, so the evidentiary layer can sit over decisions a firm is already making. Multi node attestation across fielded units is provided by Pantheon, our post-quantum Layer 1, currently on testnet, which lets independent units confirm each other's records with no central server in the middle.
Why the regulation makes this a market, not a feature
The constraints that keep data inside the walls are not going away. They are tightening. The PRA's model risk expectations under SS1/23 raise the bar for how a supervised firm governs, tests, and evidences any model that touches a material decision. UK GDPR special category rules, the NHS Data Security and Protection Toolkit, the EU AI Act's high risk classification, ITAR and EAR, the NIS Regulations, and the extraterritorial reach of the US CLOUD Act all push in the same direction. Each one narrows where regulated data can go and raises the standard of proof for what an automated system did with it.
That is why we treat sovereignty as the market rather than a checkbox. The sovereign AI market is roughly USD 40 billion in 2025 and is expected to reach about USD 148 billion by 2032. Around 0.85 million UK businesses, about fifteen per cent of the total, and roughly five million across the EU, legally cannot send their data to public cloud AI. Capital markets firms sit near the centre of that group, carrying some of the heaviest supervisory load in the economy. They are not underserved by accident. The dominant delivery model cannot lawfully reach them.
How we cover the market: the dual buyer thesis
We sell in two directions. First, we sell sovereign AI directly to the regulated firms the public cloud cannot lawfully reach, giving them capability inside their own walls with the evidence to defend it. Second, we license the patented stack to the platforms that want to reach those same firms and currently cannot. The value to a large platform is structural. A provider that adds a sovereign, on premises, air gapped layer with a verifiable audit record instantly becomes serviceable to a regulated market it is shut out of today, without asking that market to compromise on data residency or proof.
Our internal analysis maps 196 companies and 311 patent company pairs as potential licensees, including names such as Microsoft, AWS, NVIDIA, Google, Adobe, and IBM. We are precise about what that is. It is potential-licensee sizing, not a signed book and not an infringement claim. We see ourselves as an ally to the AI majors, not a challenger to them. They have the distribution and the platform relationships. We have the sovereign substrate and the filed intellectual property that lets those platforms serve buyers they cannot serve today.
The intellectual property behind the position
The moat is documented. We have 104 filed UK patent applications, roughly 2,340 claims, across 13 invention families, owned by Mickai LTD, with named inventor Mickarle Sean Junior Wagstaff-Irons. These are filed, not granted. Filing establishes priority and builds a prior-art moat, which is precisely the point at this stage: it fixes the date on the sovereign, post-quantum, audited architecture that both buyers, regulated firms and platforms alike, will need.
The company
We are Mickai LTD, a UK company, Companies House number 17166618, with manufacturing secured in Birmingham. Micky Irons is our founder and CEO. We are British by design and by intent, and we build for firms that have to answer to a regulator on Monday morning.
Our pre-seed round is opening soon, and we welcome inquiries from interested partners by email at micky@mickai.co.uk or on LinkedIn.
The bottom line for capital markets
Regulated trading firms do not have to choose between adopting artificial intelligence and honouring the rules that govern their data. We remove the trade off. The intelligence stays inside the perimeter, the audit record proves what happened without anyone having to trust the vendor, and the whole system is built to withstand supervision, litigation, and the arrival of quantum capable adversaries. That is the capability the regulated market has been waiting for, and it is live today.
Does anything leave our environment?
No. We run on your own hardware, on premises and air gapped, with zero data egress and no public cloud round trip. The models are resident inside your perimeter, and consequential actions are recorded in the Open Audit Record, which you can verify offline without any connection back to us.
How do we prove to a regulator what the system did?
Every consequential action is signed under post-quantum cryptography and hash chained into a tamper evident, append only ledger. Anyone can verify it offline, for decades, without trusting us. For a supervised firm that means the evidence for a decision is produced at the moment of the decision, not reconstructed later.
Are the patents granted?
They are filed, not granted. We have 104 filed UK patent applications, roughly 2,340 claims, across 13 invention families, owned by Mickai LTD. Filing establishes priority and builds a prior-art moat over the sovereign, post-quantum, audited architecture.